No hinting with CG should prevent outline dilation.

[bungeman, 2012-09-28 16:14:06]

Toward a solution to http://code.google.com/p/chromium/issues/detail?id=152304

[caryclark1, 2012-09-28 18:05:09]

https://codereview.appspot.com/6568065/diff/1/src/ports/SkFontHost_mac_corete...
File src/ports/SkFontHost_mac_coretext.cpp (right):

https://codereview.appspot.com/6568065/diff/1/src/ports/SkFontHost_mac_corete...
src/ports/SkFontHost_mac_coretext.cpp:751: //HACK: we do not currently support
lcd smoothed un-hinted rasterization
Note what current bug tracks this, with something like:

// FIXME: lcd smoothed un-hinted rasterization unsupported. Tracked by
http://webkit.org/b/xxx

or something like that. (WebKit mostly uses FIXME instead of HACK)

[reed1, 2012-09-28 18:18:14]

https://codereview.appspot.com/6568065/diff/1/src/ports/SkFontHost_mac_corete...
File src/ports/SkFontHost_mac_coretext.cpp (right):

https://codereview.appspot.com/6568065/diff/1/src/ports/SkFontHost_mac_corete...
src/ports/SkFontHost_mac_coretext.cpp:304: 
Can we use an enum instead, to make it clearer than a bool (esp. a bool that
begins with a negation)?

e.g.

enum GenA8Technique {
    kFromLCD_GenA8Technique,
    kFromA8_GenA8Technique
};

[bungeman, 2012-09-28 18:28:45]

https://codereview.appspot.com/6568065/diff/1/src/ports/SkFontHost_mac_corete...
File src/ports/SkFontHost_mac_coretext.cpp (right):

https://codereview.appspot.com/6568065/diff/1/src/ports/SkFontHost_mac_corete...
src/ports/SkFontHost_mac_coretext.cpp:304: 
On 2012/09/28 18:18:15, reed1 wrote:
> Can we use an enum instead, to make it clearer than a bool (esp. a bool that
> begins with a negation)?
> 
> e.g.
> 
> enum GenA8Technique {
>     kFromLCD_GenA8Technique,
>     kFromA8_GenA8Technique
> };

Negative bools are a bit of a pain. Made non-negative. Seems cleaner now, not
sure that enum is warranted.

https://codereview.appspot.com/6568065/diff/1/src/ports/SkFontHost_mac_corete...
src/ports/SkFontHost_mac_coretext.cpp:751: //HACK: we do not currently support
lcd smoothed un-hinted rasterization
On 2012/09/28 18:05:09, caryclark1 wrote:
> Note what current bug tracks this, with something like:
> 
> // FIXME: lcd smoothed un-hinted rasterization unsupported. Tracked by
> http://webkit.org/b/xxx
> 
> or something like that. (WebKit mostly uses FIXME instead of HACK) 

Yeah, that's a better message. Opened bug to track and changed comment.

[reed1, 2012-09-28 18:47:23]

lgtm

[caryclark1, 2012-09-28 19:12:00]

lgtm

[bungeman, 2012-09-28 19:24:36]

Committed revision 5728.